Development of a quantitative microbiological spoilage risk assessment (QMSRA) model for cooked ham sliced at retail.

A quantitative microbiological spoilage risk assessment model (QMSRA) for cooked ham sliced at retail was developed based on a stochastic growth model for lactic acid bacteria (LAB), which are considered as the specific spoilage organisms (SSO), and a "spoilage-response" relationship characterizing the variability in consumer's perception of spoilage. In a simulation involving 10,000 cooked ham purchases, the QMSRA model predicted a median of zero spoilage events for up to 4.5 days of storage. After storage times of 5 and 6 days, the model predicted 1,790 and 8,570 spoilage events, respectively. A sensitivity analysis showed that domestic storage temperature was the most significant factor affecting LAB concentration in cooked ham, followed by the LAB contamination level at slicing. A scenario analysis was performed testing better temperature control of consumer's refrigerators, better hygiene conditions during slicing and a combination of the two strategies. Among the tested scenarios, a 2 log reduction in the LAB contamination at slicing combined with a 2 °C decrease in domestic storage temperature resulted in zero risk of spoilage for up to 12 days of storage. The QMSRA model developed in the present study can be a useful tool for quality management decisions.

Cardinal models to describe the effect of temperature and pH on the growth of Anoxybacillus flavithermus & Bacillus licheniformis.

Anoxybacillus flavithermus and Bacillus licheniformis are among the predominant spore-formers of heat-processed foods. To our knowledge, no systematic analysis of growth kinetic data of A. flavithermus or B. licheniformis is currently available. In the present study, the growth kinetics of A. flavithermus and B. licheniformis in broth at various temperature and pH conditions were studied. Cardinal models were used to model the effect of the above-mentioned factors on the growth rates. The estimated values for the cardinal parameters Tmin,Topt,Tmax,pHmin and pH1/2 for A. flavithermus were 28.70 ± 0.26, 61.23 ± 0.16 and 71.52 ± 0.32 °C, 5.52 ± 0.01 and 5.73 ± 0.01, respectively, while for B. licheniformis they were 11.68 ± 0.03, 48.05 ± 0.15, 57.14 ± 0.01 °C, 4.71 ± 0.01 and 5.670 ± 0.08, respectively. The growth behaviour of these spoilers was also investigated in a pea beverage at 62 and 49 °C, respectively, to adjust the models to this product. The adjusted models were further validated at static and dynamic conditions and demonstrated good performance with 85.7 and 97.4% of predicted populations for A. flavithermus and B. licheniformis, respectively, being within the -10%-10% relative error (RE) zone. The developed models can be useful tools in assessing the potential of spoilage of heat-processed foods including plant-based milk alternatives.

Quantitative risk assessment of Listeria monocytogenes in ready-to-eat (RTE) cooked meat products sliced at retail stores in Greece.

A quantitative microbial risk assessment (QMRA) model predicting the listeriosis risk related to the consumption of Ready- To- Eat (RTE) cooked meat products sliced at retail stores in Greece was developed. The probability of illness per serving assessed for 87 products available in the Greek market was found highly related to the nitrite concentration; products having a lower concentration showed a higher risk per serving. The predicted 95th percentiles of the annual listeriosis cases totaled 33 of which 13 cases were <65 years old and 20 cases ≥65 years old. The highest number of cases was predicted for mortadella, smoked turkey, boiled turkey and parizer, which were the most frequently consumed product categories. Two scenarios for assessing potential interventions to reduce the risk were tested: setting a use-by date of 14 days (these products have no use-by date based on current European Union legislation) and improving the temperature control during domestic storage. The two scenarios resulted in a decrease of the 95th and 99th percentiles of the total annual cases by 97% and 88%, respectively.

Heterogeneity of single cell inactivation: Assessment of the individual cell time to death and implications in population behavior.

A direct microscopic time-lapse method, using appropriate staining for cell viability in a confocal scanning laser microscope, was used for the direct assessment of Salmonella Agona individual cell inactivation in small two-dimensional colonies exposed to osmotic stress. Individual cell inactivation times were fitted to a variety of continuous distributions using @Risk software. The best fitted distribution (LogLogistic) was further used to predict the inactivation of Salmonella populations of various initial levels using Monte Carlo simulation. The simulation results showed that the variability in inactivation kinetics is negligible for concentrations down to 100 cells and the population behavior can be described with a deterministic model. As the concentration decreases below 100 cells, however, the variability increases significantly indicating that the traditional D-value used in deterministic first order kinetic models is not valid. At a second stage, single cell behavior was monitored in larger three dimensional colonies. The results showed that colony size can affect the inactivation pattern. The effect of colony size on microbial inactivation was confirmed with validation experiments which showed a higher inactivation rate for populations consisting of single cells or small colonies compared to those consisting of cells organized in larger colonies.

Development and validation of predictive models for the effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks.

This study was undertaken to provide quantitative tools for predicting the behavior of the spoilage bacterium Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks. In the first part of the study, a growth/no growth interface model was developed, predicting the probability of growth as a function of temperature and pH. For this purpose, the growth ability of A. acidoterrestris was studied at different combinations of temperature (15-45 °C) and pH (2.02-5.05). The minimum pH and temperature where growth was observed was 2.52 (at 35 and 45 °C) and 25 °C (at pH ≥ 3.32), respectively. Then a logistic polynomial regression model was fitted to the binary data (0: no growth, 1: growth) and, based on the concordance index (98.8%) and the Hosmer-Lemeshow statistic (6.226, P = 0.622), a satisfactory goodness of fit was demonstrated. In the second part of the study, the effects of temperature (25-55 °C) and pH (3.03-5.53) on A. acidoterrestris growth rate were investigated and quantitatively described using the cardinal temperature model with inflection and the cardinal pH model, respectively. The estimated values for the cardinal parameters Tmin, Tmax, Topt and pHmin, pHmax, pHopt were 18.11, 55.68, 48.60 °C and 2.93, 5.90, 4.22, respectively. The developed models were validated against growth data of A. acidoterrestris obtained in eight commercial pasteurized fruit drinks. The validation results showed a good performance of both models. In all cases where the growth/no growth interface model predicted a probability lower than 0.5, A. acidoterrestris was, indeed, not able to grow in the tested fruit drinks; similarly, when the model predicted a probability above 0.9, growth was observed in all cases. A good agreement was also observed between growth predicted by the kinetic model and the observed kinetics of A. acidoterrestris in fruit drinks at both static and dynamic temperature conditions.

Risk assessment of fungal spoilage: A case study of Aspergillus niger on yogurt.

A quantitative risk assessment model of yogurt spoilage by Aspergillus niger was developed based on a stochastic modeling approach for mycelium growth by taking into account the important sources of variability such as time-temperature conditions during the different stages of chill chain and individual spore behavior. Input parameters were fitted to the appropriate distributions and A. niger colony's diameter at each stage of the chill chain was estimated using Monte Carlo simulation. By combining the output of the growth model with the fungus prevalence, that can be estimated by the industry using challenge tests, the risk of spoilage translated to number of yogurt cups in which a visible mycelium of A. niger is being formed at the time of consumption was assessed. The risk assessment output showed that for a batch of 100,000 cups in which the percentage of contaminated cups with A. niger was 1% the predicted numbers (median (5th, 95th percentiles)) of the cups with a visible mycelium at consumption time were 8 (5, 14). For higher percentages of 3, 5 and 10 the predicted numbers (median (5th, 95th percentiles)) of the spoiled cups at consumption time were estimated to be 24 (16, 35), 39 (29, 52) and 80 (64, 94), respectively. The developed model can lead to a more effective risk-based quality management of yogurt and support the decision making in yogurt production.

Effect of storage temperature on the lag time of Geobacillus stearothermophilus individual spores.

The lag times (λ) of Geobacillus stearothermophilus single spores were studied at different storage temperatures ranging from 45 to 59 °C using the Bioscreen C method. A significant variability of λ was observed among individual spores at all temperatures tested. The storage temperature affected both the position and the spread of the λ distributions. The minimum mean value of λ (i.e. 10.87 h) was observed at 55 °C, while moving away from this temperature resulted in an increase for both the mean and standard deviation of λ. A Cardinal Model with Inflection (CMI) was fitted to the reverse mean λ, and the estimated values for the cardinal parameters Tmin, Tmax, Topt and the optimum mean λ of G. stearothermophilus were found to be 38.1, 64.2, 53.6 °C and 10.3 h, respectively. To interpret the observations, a probabilistic growth model for G. stearothermophilus individual spores, taking into account λ variability, was developed. The model describes the growth of a population, initially consisting of N0 spores, over time as the sum of cells in each of the N0 imminent subpopulations originating from a single spore. Growth simulations for different initial contamination levels showed that for low N0 the number of cells in the population at any time is highly variable. An increase in N0 to levels exceeding 100 spores results in a significant decrease of the above variability and a shorter λ of the population. Considering that the number of G. stearothermophilus surviving spores in the final product is usually very low, the data provided in this work can be used to evaluate the probability distribution of the time-to-spoilage and enable decision-making based on the "acceptable level of risk".

Variability in the adaptive acid tolerance response phenotype of Salmonella enterica strains.

The objective of this study was the assessment of the stationary-phase, low-pH-inducible acid tolerance response (ATR) of different Salmonella enterica strains. For this purpose, 30 strains of the pathogen were grown in tryptone soy broth in the absence (non-adapted cultures) and presence (1% w/v; acid-adapted cultures) of glucose, and then subjected to 4-h acid challenge trials at pH 3.0. Surviving populations of each strain were determined at 1-h intervals, and the Weibull model was fitted to the derived microbiological data. Extensive variability in the acid stress responses of the tested S. enterica strains was observed, with the total population reductions (log CFU/ml) attained in 4 h of acid challenge ranging from 0.9 to 5.5 and from 0.6 to 7.0 for the non-adapted and acid-adapted cultures, respectively. As demonstrated by the model scale parameter δ and shape parameter p, the effect of acid adaptation on the inactivation curves was strain-specific. Although acid adaptation resulted in enhanced acid survival for the majority of the tested strains, there were strains exhibiting similar or decreased acid resistance compared to their non-adapted counterparts. Moreover, acid adaptation appeared to decrease the strain variability of δ whereas increasing the strain variability of p: the coefficient of variation of δ among the tested strains was 97.2 and 54.9% for the non-adapted and acid-adapted cultures, respectively, while the corresponding values for p were 12.7 and 48.1%. The data of the present study, which is the first one to systematically evaluate the adaptive ATR of multiple S. enterica strains, clearly demonstrate that this phenotype (attempted to be induced by growing the pathogen in the presence of glucose) is strain-dependent.

Development and application of Geobacillus stearothermophilus growth model for predicting spoilage of evaporated milk.

The presence of Geobacillus stearothermophilus spores in evaporated milk constitutes an important quality problem for the milk industry. This study was undertaken to provide an approach in modelling the effect of temperature on G. stearothermophilus ATCC 7953 growth and in predicting spoilage of evaporated milk. The growth of G. stearothermophilus was monitored in tryptone soy broth at isothermal conditions (35-67 °C). The data derived were used to model the effect of temperature on G. stearothermophilus growth with a cardinal type model. The cardinal values of the model for the maximum specific growth rate were Tmin = 33.76 °C, Tmax = 68.14 °C, Topt = 61.82 °C and µopt = 2.068/h. The growth of G. stearothermophilus was assessed in evaporated milk at Topt in order to adjust the model to milk. The efficiency of the model in predicting G. stearothermophilus growth at non-isothermal conditions was evaluated by comparing predictions with observed growth under dynamic conditions and the results showed a good performance of the model. The model was further used to predict the time-to-spoilage (tts) of evaporated milk. The spoilage of this product caused by acid coagulation when the pH approached a level around 5.2, eight generations after G. stearothermophilus reached the maximum population density (Nmax). Based on the above, the tts was predicted from the growth model as the sum of the time required for the microorganism to multiply from the initial to the maximum level ( [Formula: see text] ), plus the time required after the [Formula: see text] to complete eight generations. The observed tts was very close to the predicted one indicating that the model is able to describe satisfactorily the growth of G. stearothermophilus and to provide realistic predictions for evaporated milk spoilage.

Modelling biofilm formation of Salmonella enterica ser. Newport as a function of pH and water activity.

The effect of pH and water activity (aw) on the formation of biofilm by Salmonella enterica ser. Newport, previously identified as a strong biofilm producer, was assessed. Biofilm formation was evaluated in tryptone soy broth at 37 °C and at different combinations of pH (3.3-7.8) and aw (0.894-0.997). In total, 540 biofilm formation tests in 108 pH and aw combinations were carried out in polystyrene microtiter plates using crystal violet staining and optical density (OD; 580 nm) measurements. Since the individual effects of pH and aw on biofilm formation had a similar pattern to that observed for microbial growth rate, cardinal parameter models (CPMs) were used to describe these effects. CPMs described successfully the effects of these two environmental parameters, with the estimated cardinal values of pHmin, pHopt, pHmax, awmin and awopt being 3.58, 6.02, 9.71, 0.894 and 0.994, respectively. The CPMs assumption of the multiplicative inhibitory effect of environmental factors was validated in the case of biofilm formation using additional independent data (i.e. 430 OD data at 86 different combinations of pH and aw). The validation results showed a good agreement (r(2) = 0.938) between observed and predicted OD with no systematic error. In the second part of this study, a probabilistic model predicting the pathogen's biofilm formation boundaries was developed, and the degree of agreement between predicted probabilities and observations was as high as 99.8%. Hence, the effect of environmental parameters on biofilm formation can be quantitatively expressed using mathematical models, with the latter models, in turn, providing useful information for biofilm control in food industry environments.

Individual cell heterogeneity as variability source in population dynamics of microbial inactivation.

A statistical modeling approach was applied for describing and evaluating the individual cell heterogeneity as variability source in microbial inactivation. The inactivation data (Nt vs time) of Salmonella enterica serotype Agona, with initial concentration N0 = 10(9) CFU/ml in acidified tryptone soy broth (pH 3.5), were transformed to (N0 - Nt)/N0 vs time leading to the cumulative probability distribution of the individual cell inactivation times (ti), which was further fitted to a variety of continuous distributions using @Risk software. The best-fitted ti distribution (Gamma) was used to predict the inactivation of S. Agona populations of various N0 using Monte Carlo simulation, with the number of iterations in each simulation being equal to N0 and the number of simulations representing the variability of the population inactivation behavior. The Monte Carlo simulation results for a population with N0 = 10,000 CFU/ml showed that the variability in the predicted inactivation behavior is negligible for concentrations down to 100 cells. As the concentration decreases below 100 cells, however, the variability increases significantly. The results also indicated that the D-value used in deterministic first order kinetic models is valid only for large populations. For small populations, D-value shows a high variability, originating from individual cell heterogeneity, and, thus, can be better characterized by a probability distribution rather than a uniform value. Validation experiments with small populations confirmed the variability predicted by the statistical model. The use of the proposed approach to quantify the variability in the inactivation of mixed microbial populations, consisting of subpopulations with different probability distributions of ti, was also demonstrated.

Quantitative microbial risk assessment of Salmonella in fresh chicken patties.

Quantitative microbial risk assessment (QMRA) has witnessed rapid development within the context of food safety in recent years. As a means of contributing to these advancements, a QMRA for Salmonella spp. in fresh chicken patties for the general European Union (EU) population was developed. A two-dimensional (Second Order) Monte-Carlo simulation method was used for separating variability and uncertainty of model's parameters. The stages of industrial processing, retail storage, domestic storage, and cooking in the domestic environment were considered in the exposure assessment. For hazard characterization, a dose-response model was developed by combining 8 published dose-response models using a Pert distribution for describing uncertainty. The QMRA model predicted a mean probability of illness of 1.19*10-4 (5.28*10-5 - 3.57*10-4 95 % C.I.), and a mean annual number of illnesses per 100,000 people of 2.13 (0.96 - 6.59 95 % C.I.). Moreover, sensitivity analysis was performed, and variability in cooking preferences was found to be the most influential model parameter (r = -0.39), followed by dose-response related variability (r = 0.22), and variability in the concentration of Salmonella spp. at the time of introduction at the processing facility (r = 0.11). Various mitigation strategy scenarios were tested, from which, "increasing the internal temperature of cooking" and "decreasing shelf life" were estimated to be the most effective in reducing the predicted risk of illness. Salmonella-related illnesses exhibit particularly high severity, making them some of the most prominent zoonotic diseases in the EU. Regular monitoring of this hazard in order to further highlight its related parameters and causes is a necessary procedure. This study not only provides an updated assessment of Salmonella spp. risk associated with chicken patties, but also facilitates the identification of crucial targets for scientific investigation and implementation of real-world intervention strategies.

The impact of the proposed revised Australia's microbiological monitoring programme for beef and sheep meat exported to the EU.

The European Commission asks scientific and technical assistance from EFSA to determine the impact of the revision of the Australian monitoring programme on its ability to detect microbiological contamination. Considering that, in 2010, the European Commission determined the current Australian monitoring programme to be equivalent to the EU requirements for microbiological monitoring further to an EFSA scientific assessment, the current and proposed programmes were described and the total number of alerts was compared using a probabilistic modelling approach. In the current programme, only beef and sheep carcasses are monitored using three-class moving window sampling plans, while in the proposed programme, carcass, bulk meat, primal and offal are monitored using four two-class sampling plans and Salmonella testing is excluded. The models revealed that the current programme provides a higher number of alerts for APC, while the proposed monitoring programme provides a higher number of alerts for E. coli. For APC and E. coli combined, the mean, 5th and 95th centiles of the uncertainty distribution of the total number of alerts in the current and the proposed monitoring programme are 201 [179, 227] and 172 [149, 194] for beef, and 199 [175, 222] and 2897 [2795, 3008] for sheep, respectively. For Salmonella, there are no alerts for the proposed programme since sampling is excluded while for the current programme, the estimated mean, 5th and 95th centiles of the uncertainty distribution of the number of alerts for a 5-year period were 143 [126, 144] for heifer/steer, 1.6 [0, 4] for cow/bull and 0 [0, 0] for lamb/sheep. Overall, for APC and E. coli, the estimated total number of alerts was similar (beef) or higher (sheep) for the proposed compared to the current programme. In contrast, Salmonella sampling is excluded from the proposed programme and thus cannot detect the number of current alerts.

BSE risk posed by ruminant collagen and gelatine derived from bones.

The European Commission requested an estimation of the BSE risk (C-, L- and H-BSE) from gelatine and collagen derived from ovine, caprine or bovine bones, and produced in accordance with Regulation (EC) No 853/2004, or Regulation (EC) No 1069/2009 and its implementing Regulation (EU) No 142/2011. A quantitative risk assessment was developed to estimate the BSE infectivity, measured in cattle oral infectious dose 50 (CoID50), in a small size batch of gelatine including one BSE-infected bovine or ovine animal at the clinical stage. The model was built on a scenario where all ruminant bones could be used for the production of gelatine and high-infectivity tissues remained attached to the skull (brain) and vertebral column (spinal cord). The risk and exposure pathways defined for humans and animals, respectively, were identified. Exposure routes other than oral via food and feed were considered and discussed but not assessed quantitatively. Other aspects were also considered as integrating evidence, like the epidemiological situation of the disease, the species barrier, the susceptibility of species to BSE and the assumption of an exponential dose-response relationship to determine the probability of BSE infection in ruminants. Exposure to infectivity in humans cannot be directly translated to risk of disease because the transmission barrier has not yet been quantified, although it is considered to be substantial, i.e. much greater amounts of infectivity would be needed to successfully infect a human and greater in the oral than in the parenteral route of exposure. The probability that no new case of BSE in the cattle or small ruminant population would be generated through oral exposure to gelatine made of ruminant bones is 99%-100% (almost certain) This conclusion is based on the current state of knowledge, the epidemiological situation of the disease and the current practices, and is also valid for collagen.

Evaluation of alternative methods of tunnel composting (submitted by the European Composting Network) II.

Two alternative methods for producing compost in a tunnel, from certain category (Cat.) 3 animal by-products (ABP) and other non-ABP material, were assessed. The first method proposed a minimum temperature of 55°C for 72 h and the second 60°C for 48 h, both with a maximum particle size of 200 mm. The assessment of the Panel on Biological Hazards (BIOHAZ) exclusively focused on Cat. 3 ABP materials (catering waste and processed foodstuffs of animal origin no longer intended for human consumption). The proposed composting processes were evaluated for their efficacy to achieve a reduction of at least 5 log10 of Enterococcus faecalis and Salmonella Senftenberg (775W, H2S negative) and at least 3 log10 of relevant thermoresistant viruses. The applicant provided a list of biological hazards that may enter the composting process and selected parvoviruses as the indicator of the thermoresistant viruses. The evidence provided by the applicant included: (a) literature data on thermal inactivation of biological hazards; (b) results from validation studies on the reduction of E. faecalis, Salmonella Senftenberg 775W H2S negative and canine parvovirus carried out in composting plants across Europe; (c) and experimental data from direct measurements of reduction of infectivity of murine parvovirus in compost material applying the time/temperature conditions of the two alternative methods. The evidence provided showed the capacity of the proposed alternative methods to reduce E. faecalis and Salmonella Senftenberg 775W H2S negative by at least 5 log10, and parvoviruses by at least 3 log10. The BIOHAZ Panel concluded that the two alternative methods under assessment can be considered to be equivalent to the processing method currently approved in the Commission Regulation (EU) No 142/2011.

Update of the list of qualified presumption of safety (QPS) recommended microbiological agents intentionally added to food or feed as notified to EFSA 19: Suitability of taxonomic units notified to EFSA until September 2023.

The qualified presumption of safety (QPS) process was developed to provide a safety assessment approach for microorganisms intended for use in food or feed chains. The QPS approach is based on an assessment of published data for each taxonomic unit (TU), with respect to its taxonomic identity, the body of relevant knowledge and safety concerns. Safety concerns identified for a TU are, where possible, confirmed at the species/strain or product level and reflected by 'qualifications'. In the period covered by this Statement, no new information was found that would change the status of previously recommended QPS TUs. Of 71 microorganisms notified to EFSA between April and September 2023 (30 as feed additives, 22 as food enzymes or additives, 7 as novel foods and 12 from plant protection products [PPP]), 61 were not evaluated because: 26 were filamentous fungi, 1 was Enterococcus faecium, 5 were Escherichia coli, 1 was a bacteriophage (all excluded from the QPS evaluation) and 28 were TUs that already have a QPS status. The other 10 notifications belonged to 9 TUs which were evaluated for a possible QPS status: Ensifer adhaerens and Heyndrickxia faecalis did not get the QPS recommendation due to the limited body of knowledge about their occurrence in the food and/or feed chains and Burkholderia ubonensis also due to its ability to generate biologically active compounds with antimicrobial activity; Klebsiella pneumoniae, Serratia marcescens and Pseudomonas putida due to safety concerns. K. pneumoniae is excluded from future QPS evaluations. Chlamydomonas reinhardtii is recommended for QPS status with the qualification 'for production purposes only'; Clostridium tyrobutyricum is recommended for QPS status with the qualification 'absence of genetic determinants for toxigenic activity'; Candida oleophila has been added as a synonym of Yarrowia lipolytica. The Panel clarifies the extension of the QPS status for genetically modified strains.

Scientific Committee guidance on appraising and integrating evidence from epidemiological studies for use in EFSA's scientific assessments.

EFSA requested its Scientific Committee to prepare a guidance document on appraising and integrating evidence from epidemiological studies for use in EFSA's scientific assessments. The guidance document provides an introduction to epidemiological studies and illustrates the typical biases, which may be present in different epidemiological study designs. It then describes key epidemiological concepts relevant for evidence appraisal. This includes brief explanations for measures of association, exposure assessment, statistical inference, systematic error and effect modification. The guidance then describes the concept of external validity and the principles of appraising epidemiological studies. The customisation of the study appraisal process is explained including tailoring of tools for assessing the risk of bias (RoB). Several examples of appraising experimental and observational studies using a RoB tool are annexed to the document to illustrate the application of the approach. The latter part of this guidance focuses on different steps of evidence integration, first within and then across different streams of evidence. With respect to risk characterisation, the guidance considers how evidence from human epidemiological studies can be used in dose-response modelling with several different options being presented. Finally, the guidance addresses the application of uncertainty factors in risk characterisation when using evidence from human epidemiological studies.

Persistence of microbiological hazards in food and feed production and processing environments.

Listeria monocytogenes (in the meat, fish and seafood, dairy and fruit and vegetable sectors), Salmonella enterica (in the feed, meat, egg and low moisture food sectors) and Cronobacter sakazakii (in the low moisture food sector) were identified as the bacterial food safety hazards most relevant to public health that are associated with persistence in the food and feed processing environment (FFPE). There is a wide range of subtypes of these hazards involved in persistence in the FFPE. While some specific subtypes are more commonly reported as persistent, it is currently not possible to identify universal markers (i.e. genetic determinants) for this trait. Common risk factors for persistence in the FFPE are inadequate zoning and hygiene barriers; lack of hygienic design of equipment and machines; and inadequate cleaning and disinfection. A well-designed environmental sampling and testing programme is the most effective strategy to identify contamination sources and detect potentially persistent hazards. The establishment of hygienic barriers and measures within the food safety management system, during implementation of hazard analysis and critical control points, is key to prevent and/or control bacterial persistence in the FFPE. Once persistence is suspected in a plant, a 'seek-and-destroy' approach is frequently recommended, including intensified monitoring, the introduction of control measures and the continuation of the intensified monitoring. Successful actions triggered by persistence of L. monocytogenes are described, as well as interventions with direct bactericidal activity. These interventions could be efficient if properly validated, correctly applied and verified under industrial conditions. Perspectives are provided for performing a risk assessment for relevant combinations of hazard and food sector to assess the relative public health risk that can be associated with persistence, based on bottom-up and top-down approaches. Knowledge gaps related to bacterial food safety hazards associated with persistence in the FFPE and priorities for future research are provided.

Public health aspects of Vibrio spp. related to the consumption of seafood in the EU.

Vibrio parahaemolyticus, Vibrio vulnificus and non-O1/non-O139 Vibrio cholerae are the Vibrio spp. of highest relevance for public health in the EU through seafood consumption. Infection with V. parahaemolyticus is associated with the haemolysins thermostable direct haemolysin (TDH) and TDH-related haemolysin (TRH) and mainly leads to acute gastroenteritis. V. vulnificus infections can lead to sepsis and death in susceptible individuals. V. cholerae non-O1/non-O139 can cause mild gastroenteritis or lead to severe infections, including sepsis, in susceptible individuals. The pooled prevalence estimate in seafood is 19.6% (95% CI 13.7-27.4), 6.1% (95% CI 3.0-11.8) and 4.1% (95% CI 2.4-6.9) for V. parahaemolyticus, V. vulnificus and non-choleragenic V. cholerae, respectively. Approximately one out of five V. parahaemolyticus-positive samples contain pathogenic strains. A large spectrum of antimicrobial resistances, some of which are intrinsic, has been found in vibrios isolated from seafood or food-borne infections in Europe. Genes conferring resistance to medically important antimicrobials and associated with mobile genetic elements are increasingly detected in vibrios. Temperature and salinity are the most relevant drivers for Vibrio abundance in the aquatic environment. It is anticipated that the occurrence and levels of the relevant Vibrio spp. in seafood will increase in response to coastal warming and extreme weather events, especially in low-salinity/brackish waters. While some measures, like high-pressure processing, irradiation or depuration reduce the levels of Vibrio spp. in seafood, maintaining the cold chain is important to prevent their growth. Available risk assessments addressed V. parahaemolyticus in various types of seafood and V. vulnificus in raw oysters and octopus. A quantitative microbiological risk assessment relevant in an EU context would be V. parahaemolyticus in bivalve molluscs (oysters), evaluating the effect of mitigations, especially in a climate change scenario. Knowledge gaps related to Vibrio spp. in seafood and aquatic environments are identified and future research needs are prioritised.